MARINE ECOLOGY PROGRESS SERIES Mar Ecol Prog Ser
Vol. 209: 257–273, 2001
Published January 5
Methods for estimating decapod larval supply and settlement: importance of larval behavior and development stage Per-Olav Moksnes*, Håkan Wennhage Göteborg University, Kristineberg Marine Research Station, 45034 Fiskebäckskil, Sweden
ABSTRACT: In marine benthic organisms with a pelagic larval phase, assessment of recruitment regulation necessitates accurate estimates of larval supply and initial settlement densities. We assessed 2 commonly used methods for estimating decapod larval supply, plankton net tows and artificial settlement substrates (ASS), together with a new technique using passive-migration traps. The aim was to evaluate how these methods estimate larval supply and settlement of 3 common decapod species (the shore crab Carcinus maenas L., the brown shrimp Crangon crangon L. and the grass shrimps Palaemon spp.) in a microtidal nursery area on the Swedish west coast, and to assess how these estimates relate to the juvenile recruitment of these species. Discrete plankton net tows outside the shallow nursery area collected a higher proportion of larvae at an early development stage, compared to the other methods, and produced only a snapshot of the variable water-column abundance of shore crab and grass shrimp larvae that correlated poorly with numbers collected with ASS at the same location. Artificial settlement substrates appeared to produce good, integrated relative estimates of shore crab larval supply and settlement. The number of larvae collected from ASS correlated significantly with larval abundance estimated by passive-migration traps in shallow nursery areas, and reflected changes in shore crab settlement densities in caged mussel habitats. We did not detect any effect of predation on ASS placed in nursery areas. However, a large proportion (an estimated 83%) of shore crab postlarvae appeared to emigrate from ASS immersed for periods longer than 12 h, possibly during the night. All grass shrimp larvae appeared to emigrate from the collectors within 24 h, suggesting that ASS immersed longer than 12 h do not produce useful integrated estimates of larval supply for this species. These results demonstrate the importance of assessing postlarval emigration patterns from ASS to optimize immersion and collection time, and to avoid confounded estimates of larval abundance and settlement. A new method using replicated passive migration traps that fished continuously in opposite directions (on-shore and off-shore) gave promising integrated estimates of net fluxes (total number immigrating minus number emigrating per unit time) of both pelagic postlarvae and early benthic juvenile stages of crabs and shrimp. The majority of the brown shrimp recruits were young juveniles, demonstrating the importance of incorporating juvenile migration in recruitment studies of motile benthic species. High numbers of shore crab and grass shrimp larvae were found to emigrate from the bay, indicating that many decapod postlarvae found in nursery areas may be transitional. The emigrating larvae were on average in an earlier development stage than those migrating to the bay. These short-term experiments demonstrate the importance of assessing larval development stage and migratory behavior when estimating larval supply and settlement for recruitment studies. KEY WORDS: Larval supply · Settlement · Juvenile migration · Development stage · Sampling problems Resale or republication not permitted without written consent of the publisher
INTRODUCTION The relative importance of pre- and post-settlement processes for population regulation in benthic organ*E-mail:
[email protected] © Inter-Research 2001
isms is presently a major focus in marine ecology (Caley et al. 1996). To assess settler-recruit relationships or density-dependent post-settlement mortality in a population, it is imperative to obtain an accurate estimate of initial settlement densities. In nature this is very difficult to measure, because larvae are com-
258
Mar Ecol Prog Ser 209: 257–273, 2001
monly small and cryptic, and also because mortality can occur during settlement. One way to overcome these difficulties is to measure the abundance of postlarvae in the water column above the benthic population as a substitute for settlement densities. However, recent studies imply that discrete estimates of larval abundance in the water column, collected with plankton nets or pumps (the traditional methods) fail to take into account the high spatio-temporal variability in larval abundance and flow velocity (Gaines & Bertness 1993) and behavioral aspects of the larvae (Miron et al. 1995), and therefore pose a serious impediment to progress in this field. Moreover, postlarval densities in the water may not reflect settlement densities if the larvae are not competent to settle. Quantitative measurements of postlarval development state may therefore be necessary to relate planktonic availability with settlement patterns (Lipcius et al. 1990). An alternative method for estimating larval supply and settlement densities is to measure larval and juvenile abundance on artificial substrata where mortality is thought to be low. Artificial settlement substrates (ASS) or collectors that sample passively over an immersion period have been used for over 20 yr to estimate larval abundance and settlement in decapod species (i.e. the western rock lobster Panulirus longipes cygnus [Phillips 1972], the grapsid crab Pachygrapsus crassipes [Shanks 1983], the rock crab Cancer irrotatus [Beniger et al. 1986], the blue crab Callinectes sapidus [Goodrich et al. 1989, van Montfrans et al. 1990], the red king crab Paralithodes camtschaticus [Blau & Byersdorfer 1994], and the Dungeness crab Cancer magister [Eggleston & Armstrong 1995]). This method was recently standardized to quantify daily settlement patterns of blue crab postlarvae over an extensive geographical area covering both the Atlantic and the Gulf coasts of the USA (van Montfrans et al. 1995 and references therein). The ASS method makes use of the thigmotactic behavior of many decapod postlarvae to cling to encountered objects (Goodrich et al. 1989), and therefore integrates larval behavior and abundance in a relative measurement of larval supply. In comparison with postlarvae collected with plankton nets, ASS catch a higher proportion of postlarvae in a late developmental stage (i.e. in premolt; Lipcius et al. 1990), and may therefore give better estimates of settlement densities. Since this technique samples continuously, temporal variation in larval abundance and flow velocity are of lesser concern. One potential problem with passive collectors is that postlarval densities on ASS will not reflect settlement densities if larvae that are not competent to settle also cling to the collectors. If many larvae cling only temporarily to the ASS, estimates of larval abundance may
also be confounded, since the rate of emigration may vary in time and space as a result of light conditions, tidal phase, or larval density on the collector. The possibility that the ASS may only be a transitional substrata for postlarvae has rarely been reflected upon in earlier studies (but see Goodrich et al. 1989, Olmi et al. 1990) and has, to our knowledge, never been investigated. Similarly, the potential problem of variable predation on collectors has been little discussed and, as far as we know, never been assessed. In motile species, post-settlement dispersal by juveniles may also contribute to the juvenile recruitment of local populations (see Palmer et al. 1996 for review). If a large proportion of the recruitment occurs through juvenile migration, estimates of larval supply and settlement densities alone will not suffice to assess recruitment regulation. Therefore, it is important to understand the significance of post-settlement migration when using estimates of settlement in recruitment studies. Still, juvenile migration is rarely incorporated into studies of recruitment regulation, and there is a general lack of methods that quantitatively assess this process. This study, based on 3 short field experiments, assessed 3 different methods for estimating larval supply and settlement in decapod species: (1) plankton net tows, (2) artificial settlement substrates, and (3) a new method using passive-migration traps to estimate fluxes of both pelagic larvae and benthic juveniles to shallow nursery areas. The aim was to evaluate each sampling method to be used in recruitment studies of 3 decapod species (the shore crab Carcinus maenas L., the brown shrimp Crangon crangon L. and the grass shrimps Palaemon adspersus and P. elegans Rathke). Specifically, the aim was to assess how different relative estimates of larval supply and settlement were affected by temporal variation in larval abundance and flow velocities, development stage of the larvae, migratory behavior of the larvae, and predation. A final goal was to assess the importance of post-settlement migration for juvenile recruitment in order to evaluate if this process needs to also be incorporated in recruitment studies of these species.
MATERIALS AND METHODS Three field experiments were conducted during the summer of 1995 and 1996 in the Gullmarsfjord area on the Swedish west coast (58° 15’ N, 11° 27’ E), which has a small tidal amplitude of less than 30 cm. During the summer and fall months, high densities of postlarvae and juveniles of shore crab Carcinus maenas, brown shrimp Crangon crangon and grass shrimps Palaemon spp. are found in shallow, soft sediment ( 0.05
Fig. 3. Expt 1. Carcinus maenas. Mean number of shore crabs (megalopae and first-instar crabs; + SE; n = 3) collected from artificial settlement substrates (ASS) immersed for different periods of time (12 to 96 h). (a) ASS exposed for 12 h during the day (09:00 to 21:00 h); and the night (21:00 to 09:00 h); (b) ASS exposed for 24 h and the sum of the concurrently fished 12 h ASS over 24 h; (c) ASS exposed for 48 h and the sum of the concurrently fished 12 and 24 h ASS over 48 h; (d) ASS exposed for 96 h and the sum of the concurrently fished 12, 24 and 48 h ASS over 96 h. Shaded areas represent the proportion of first-instar crabs
264
Mar Ecol Prog Ser 209: 257–273, 2001
period. This difference was not found for the 48 h ASS, where the number of shrimp appeared to decrease during the second half (apparently reflecting the 12 h ASS night sample; Fig. 4a,c), causing a significant interaction effect (ANOVA: F = 6.0, df = 2,12, p < 0.05; Fig. 4c). The 3 immersion times all differed significantly for the second period, whereas no differences were detected for the first period (SNK test, p < 0.05). The total number of shrimp over 96 h was significantly higher on the 12 h (230 shrimp) and 24 h ASS (160 shrimp) than for the 2 longer immersion times (48 h: 63 shrimp; 96 h: 22 shrimp), which did not differ significantly from each other (ANOVA: F = 15.7, df = 3, 8, p < 0.001; SNK test at p < 0.05; Fig. 4d).
Expt 2. Benthic-pelagic migration trap — a pilot study The wind conditions during Expt 2 were variable, with on- and offshore winds between 1 and 14 m s–1 (Fig. 5a). The average tidal amplitude was 20 cm, and an incoming high-pressure system caused a 40 cm drop in mean water level half way through the experiment (Fig. 5b). Current velocities measured at the opening of the bay at 1 m below the surface above 3 m depth varied between close to zero and 14 cm s–1 (Fig. 5c), and showed a relatively constant current direction, varying between 300 and 360°, and being parallel with the bay opening most of the time. Drift markers indicated that the surface current direction inside the bay reflected the wind direction at most times. No effect of the tides was seen in the current Table 2. Expt 1. Palaemon spp. Number of grass shrimp larvae as a function of immersion time (12 and 24 h) and date (August 8 to 12) over 24 h periods. Two-fixed-factor ANOVA model Source of variation
SS
df
MS
F
Immersion time (A) Date (B) A×B Error
1882 9428 2195 7053
1 3 3 16
1882 3143 732 441
4.3 (ns) 7.1** 1.6 (ns)
**p > 0.01; ns p > 0.05
Fig. 4. Expt 1. Palaemon spp. Mean number of grass shrimp larvae (+ SE; n = 3) collected from artificial settlement substrates (ASS) immersed for different periods of time (12 to 96 h). (a) ASS exposed for 12 h during the day (09:00 to 21:00 h) and the night (21:00 to 09:00 h); (b) ASS exposed for 24 h and the sum of the concurrently fished 12 h ASS over 24 h; (c) ASS exposed for 48 h and the sum of the concurrently fished 12 and 24 h ASS over 48 h; (d) ASS exposed for 96 h and the sum of the concurrently fished 12, 24 and 48 h ASS over 96 h
Moksnes & Wennhage: Estimating larval supply and settlement
data (Fig. 5b,c). Temperature and salinity of the surface water varied between 15 and 17°C, and 26 and 30 PSU, respectively. The initial assessment of the MT in July showed that this new method could catch high numbers of both planktonic larvae and migrating juveniles under a variety of different wind and current conditions, and that the estimate of larval and juvenile fluxes had a high enough precision to detect temporal differences and significant net fluxes. A significant net in-flux of both shore crab megalopae (no first-instar crabs were found) and brown shrimp larvae and juveniles to the shallow area of the bay inside the traps were found during the 11 d period (Table 3, Fig. 5d,e). The abundance of both species showed significant variation between days, but with higher numbers in the MT-in than in the MT-out on all days (with the apparent exception of shore crabs on the last day: Fig. 5d,e). High numbers of shore crab megalopae were consistently collected in the MT-out, approximately 45% on average of the number coming in. In contrast, few brown shrimp (< 4% on average) were collected in the MT that was fishing outgoing water. Immigrating brown shrimp consisted of an equal proportion of postTable 3. Expt 2. Carcinus maenas and Crangon crangon. Number of shore crab megalopae and brown shrimp [log (x + 1)-transformed] as a function of migration-trap catch direction (in and out) and date (July 14 to 25). Two-fixedfactor ANOVA models Source of variation Shore crab Direction (A) Date (B) A×B Error Brown shrimp Direction (A) Date (B) A×B Error
SS
df
MS
F
1.75 5.16 2.25 2.67
1 10 10 22
1.75 0.52 0.23 0.12
14.4*** 4.3** 1.9 (ns)
15.48 1.89 0.71 1.32
1 10 10 22
15.48 0.19 0.07 0.06
258**** 3.1* 1.2 (ns)
****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05; ns p > 0.05
Fig. 5. Expt 2. Physical and biological data from Bökevik from July 14 to 25, 1996. (a) Wind speed and direction (average over 3 h periods), bars follow the wind direction in relation to offshore (out; 32°) and onshore (in; 212°) of the bay; (b) water level (continuous measurements); (c) current velocity (averaged over 1 h periods) at 1 m depth at the opening of the bay above 3 m depth; (d) Carcinus maenas: mean number of shore crab megalopae (± SE; n = 2) collected with benthic-pelagic migration traps (MTs) located inside the bay and facing offshore (MT-in) or onshore (MT-out) during 24 h periods; (e) Crangon crangon: mean number of brown shrimp (postlarvae and juveniles; ± SE; n = 2) collected with the MT over 24 h periods
265
larvae (i.e. 1.0 to 1.2 mm CL; Criales & Anger 1986) and young juveniles (1.3 to 2.5 mm CL). The average size of the brown shrimp collected in the MT-in and MT-out were similar (1.3 and 1.5 mm CL, respectively).
266
Mar Ecol Prog Ser 209: 257–273, 2001
Expt 3. Assessment of 3 different methods for estimates of decapod larval supply and settlement During Expt 3, a continuous offshore wind of variable strength (2 to 14 m s–1) was present (Fig. 6a). Average tidal amplitude was 29 cm during this period (Fig. 6b). Current measurements at 1 m below the surface above 3 m depth at the opening of the bay showed a relatively constant current velocity between 1 and 4 cm s–1 at most times (Fig. 6c), but a highly variable direction (between 150 and 360°), that did not reflect the relatively stable wind conditions. In contrast, the current direction at the surface inside the bay, measured with drift markers and passively floating markers, appeared to always follow the constant outward wind direction. The tidal changes were not reflected in the current data (Fig. 6b,c). Temperature and salinity in the surface water varied between 17 and 18°C, and 24 and 25 PSU, respectively.
Shore crab larval supply During the 6 d evaluation of different methods for estimates of larval supply, the bay experienced a peak in abundance of megalopae, and several methods produced strikingly different results (Fig. 6d). The number of shore crab collected from the ASS fished just outside the bay increased significantly during the first 3 d from 52 crabs (megalopae and first-instar crabs) ASS–1 24 h–1 to a major peak of 793 crabs ASS–1 24 h–1 on August 2 (ANOVA: F = 6.6, df = 5,12, p < 0.01; SNK test, p < 0.05; Fig. 6d). During the same period the megalopae collected with the plankton net at the same location decreased to close to zero. The number of megalopae increased significantly in the plankton net tows during the last 3 d (up to 206 megalopae per tow; approximately 2 megalopae m– 3; ANOVA: F = 50.0, df = 5,12, p < 0.0001; SNK test, p < 0 .05; Fig. 6d), when high numbers were still collected on the ASS; however,
Fig. 6. Expt 3. Physical (a, b) and shore crab (d, e, f) data from Bökevik from July 30 to August 5, 1996. (a) Wind speed and direction; (b) water level; (c) current velocity (see Fig. 5 legend for details of a, b, c). (d) Mean number of Carcinus maenas (megalopae and first-instar crabs; ± SE; n = 3) collected with 4 different methods: plankton net tows at the opening of the bay, artificial settlement substrates located at the opening of the bay (ASS-outside), ASS located inside the bay, and benthic-pelagic migration traps (MTs) located inside the bay and facing offshore (MT-in); (e) number collected with ASS located inside the bay at 0.6 m depth with or without predation-protective caging; (f) number collected with MTs located inside the bay facing offshore (MT-in) or onshore (MT-out). Note that data from ASS-inside and MT-in are shown in 2 different graphs for visual clarity
267
Moksnes & Wennhage: Estimating larval supply and settlement
Table 4. Expt 3. Carcinus maenas. Number of shore crab settlers [log(x +1)-transformed] sampled with artificial settlement substrate (ASS) as a function of cage-treatment (cage, no cage) and date (July 30 to August 5), and numbers of settlers sampled with migration traps (MT) as a function of direction (in, out) and date. Two-fixed-factor ANOVA models Source of variation ASS Cage (A) Date (B) A×B Error MT Direction (A) Date (B) A×B Error
MS
F
SS
df
0.001 2.71 0.31 0.55
1 5 5 24
0.001 0.001 (ns) 0.54 9.7**** 0.01 0.50 (ns) 0.008
0.012 6.62 0.23 1.17
1 5 5 24
0.012 1.33 0.05 0.05
0.24 (ns) 27.2**** 0.94 (ns)
Fig. 7b). The megalopae collected in the plankton net tows showed significantly longer TTM (4.4 d) than all the other methods (Fig. 7b). Megalopae collected on ASS-outside showed significantly longer TTM (3.4 d) than ASS-inside (2.5 d). A trend towards longer TTM among megalopae from MT-out (3.0 d) compared to MT-in (2.6 d) could be seen (Fig. 7b), but was not significant (SNK test, p < 0.05).
Shore crab settlement Settlement of shore crabs (megalopae and first-instar crabs) on natural blue mussel habitat in cages increased significantly from 112 crabs per mussel patch the first 48 h period to 354 crabs per mussel patch the
****p < 0.0001; ns p > 0.05
there was no correlation between the 2 methods (correlation analysis: r = 0.08, p = 0.88). Inside the bay, the number of megalopae on ASS was in general lower (39 to 262 crabs ASS–1 24 h–1), and reached a significant peak 1 to 2 d after the peak on ASS outside the bay (Table 4; SNK test, p < 0.05; Fig. 6d), resulting in a poor correlation between ASS-outside and ASS-inside (r = 0.19, p = 0.71). No significant difference was found between the ASS-inside and ASS-cages (Table 4), and there was a significant correlation between the 2 treatments (r = 0.90, p = 0.015; Fig. 6e). High numbers of crabs were collected in the MTs (59 to 1260 crabs MT–1 24 h–1). with a significant peak during the last 3 d, but no significant difference was found between the 2 directions on any date during this period (Table 4; SNK test, p < 0.05; Fig. 6f), precluding estimates of net fluxes. The MT-in and ASS-inside correlated significantly over the 6 d (r = 0.84, p = 0.035).
Shore crab developmental stage Both the proportion of first-instar crabs in the samples, and the metamorphosis rates of collected megalopae indicated that the different methods caught postlarvae of different development stages. The proportion of first-instar crabs was significantly higher in the MTin (17%) than in the MT-out (12%), which in turn were both significantly higher than the ASS methods (ASSoutside = 1.6%, ASS-inside = 2.5%, ASS-cage = 3.2%); the ASS methods did not differ from each other (Fig. 7a, Table 5). Only 1 first-instar crab was found in the plankton tows. In the laboratory experiment, significant differences in time to metamorphosis (TTM) was found between the different methods (Table 5,
Plankton net tows
ASSoutside
ASSinside
ASS-cage
MT-in
MT-out
Fig. 7. Expt 3. Carcinus maenas. (a) Mean proportion of firstinstar shore crabs (+ SE; n = 3) collected with 6 different methods: plankton net tows at the opening of the bay, artificial settlement substrates located at the opening of the bay (ASS-outside), ASS located inside the bay, ASS located inside the bay with predation-protective caging, benthic pelagic migration traps located inside the bay and facing the opening (MT-in), and MT facing the back of the bay (MT-out) in Bökevik from July 30 to August 5, 1996 (data are pooled for the 6 d); (b) mean time to metamorphosis (TTM; + SE; n = 2) for shore crab megalopae collected with 5 different methods in Bökevik July 30 to August 5 (TTM was not investigated for ASS-cage)
268
Mar Ecol Prog Ser 209: 257–273, 2001
Table 5. Expt 3. Carcinus maenas. Proportion of first-instar shore crabs (square-root-transformed) as a function of method (plankton net, ASS-outside, ASS-inside, ASS-cage, MTin and MT-out) and date (July 30 to August 5), and time to metamorphosis as a function of method (plankton net, ASS-outside, ASS-inside, MT-in and MT-out) and catch date (August 2 to 4). Two-fixed-factor ANOVA models Source of variation Proportion 1st instars Method (A) Date (B) A×B Error Metamorphosis rate Method (A) Date (B) A×B Error
SS
df
MS
F
1.84 0.24 0.31 0.55
5 5 25 72
0.37 0.05 0.01 0.008
48.1**** 6.4** 1.6 (ns)
13.16 5.78 3.22 2.55
4 2 8 14
3.29 2.89 0.40 0.18
17.6**** 15.6*** 2.2 (ns)
****p < 0.0001; ***p < 0.001; **p < 0.01; ns p > 0.05
Table 6. Expt 3. Carcinus maenas. Sum of shore crabs [log(x +1)-transformed] after 48 h as a function of method (cage, ASS-outside, ASS-inside, ASS-cage, MT-in) and period (July 30 to August 1, and August 1 to 3). Two-fixed-factor ANOVA model Source of variation Method (A) Period (B) A×B Error
SS
df
MS
F
1.50 1.30 0.37 0.61
3 1 3 14
0.50 1.30 0.12 0.04
11.4*** 29.8**** 2.8 (ns)
outside were relatively low (1 to 17 shrimp ASS–1 24 h–1), and showed a significant increase the last 3 d (ANOVA: F = 4.5, df = 5,12, p < 0.05; SNK test, p < 0.05), but no significant correlation with the plankton net tows (r = 0.64, p = 0.17). Inside the bay the number of grass shrimp was low (MT-in: 3 to 9 shrimp 24 h–1; ASS-inside: 2 to 4 shrimp 24 h–1) and showed no significant differences between any dates or methods, nor any significant correlation between methods (all p > 0.6). The grass shrimp larvae consisted of both postlarvae and Larval Stage V, and the proportion of postlarvae varied significantly between methods (ANOVA: F = 14.8, df = 5, 30, p < 0.0001; Fig. 9b). A low proportion of postlarvae was collected in the plankton tows (1%) which was similar to the proportion collected in the MT-out (12%) but significantly lower than the proportion collected with all other methods (41 to 94%). MTin collected a significantly higher proportion of postlarvae (63%) than MT-out. A similar proportion of postlarvae was found on ASS-outside and ASS-inside (41 and 52%, respectively), whereas a significantly higher proportion of postlarvae was found in the ASScage (94%; SNK test at p < 0.01). No juvenile grass shrimp were collected with any method.
****p < 0.0001; ***p < 0.001; ns p > 0.05
second 48 h period. The proportion of first-instar crabs collected in the cages was similar in the 2 periods, 20 and 18%, respectively. A corresponding significant increase in larval supply was detected with all 4 methods assessed (Table 6, Fig. 8).
Grass shrimps Similarly to the shore crab, the measurements of grass shrimp larvae showed strikingly different patterns between the locations inside and outside the bay, and different methods appeared to catch larvae at different development stages (Fig. 9a,b). High numbers of grass shrimp larvae were collected in plankton net tows, and showed a significant variation between days, with a major peak (217 shrimp tow–1) on the final day (ANOVA: F = 16.7, df = 5,12, p < 0.0001; SNK test, p < 0.05). During the same period, the numbers on ASS-
Cage
ASS-outside ASS-inside
ASS-cage
MT-in
Fig. 8. Expt 3. Carcinus maenas. Mean number of shore crabs (megalopae and first-instar crabs; + SE; n = 2 or 3) collected in Bökevik during two 48 h periods with 5 different methods: natural blue mussel habitat provided with predation-protective caging (collected after 48 h), artificial settlement substrates located at the opening of the bay (ASS-outside), ASS located inside the bay, ASS located inside the bay with predation-protective caging (ASS-cage), and benthic-pelagic migration traps located inside the bay and facing offshore (MT-in)
Moksnes & Wennhage: Estimating larval supply and settlement
269
Brown shrimp No brown shrimp Crangon crangon were collected with either plankton net tows or ASS, even though high numbers (up to 30 shrimp MT–1 24 h–1) were collected in the MT-in. The number of brown shrimp collected in MT-in was significantly higher the first and the last 2 d compared to the other days and compared to MT-out, where numbers were always low (on average 14% of the number collected in MT-in), causing a significant interaction effect (ANOVA: F = 7.78, df = 5, 24, p < 0.001; SNK test, p < 0.05, Fig. 9c). Immigrating brown shrimp were dominated by young juveniles (1.3 to 3.0 mm CL), and few postlarvae (10-fold difference in abundance between days, as estimated with both the MT and the ASS) and did not correlate to the generally constant flow-velocity (1 to 4 cm s–1) during Expt 3, suggesting a patchy distribution of the planktonic megalopae. Larval patchyness produces a high temporal variation in larval water-column abundance and poses a severe problem for methods that involve discrete sampling. In this study, the continuously fishing ASS-outside showed a peak of megalopal abundance that was not found in the discrete plankton net samples for several days, and no correlation was found between larval abundance estimated by the 2 methods. The problem of high temporal variation in larval abundance may have been accentuated by the unpredictable current regimes in this microtidal area. The lack of correlation between current velocities and tidal amplitude makes it difficult to determine when to sample, and calls for very frequent or continuous sampling if this method is to be used successfully. Limitations with discrete plankton net tows have also been demonstrated in tidal areas for estimating larval supply of intertidal barnacles (Gaines & Bertness 1993). Furthermore, analysis of the metamorphosis rate of the sampled shore crab postlarvae demonstrated that megalopae collected with plankton nets outside the bay were in an earlier development stage than the megalopae collected with the other methods. Similarly, grass shrimp larvae collected with plankton tows consisted almost entirely of Stage V larvae, and correlated poorly with the number collected on the ASS and to larvae entering the bay, which were dominated by the postlarval stage. Thus, the water-column abundance of larvae estimated with discrete plankton tows may not reflect settlement densities since all larvae may not be competent to settle. Our results are consistent with studies on blue crabs in which megalopae collected with plankton nets showed an earlier development stage than megalopae collected with ASS (Lipcius et al. 1990), and a poor correlation in abundance between the 2 methods (Olmi et al. 1990). Finally, daytime surface plankton net tows were unsuccessful in collecting brown shrimp larvae and juveniles, even though high numbers were concurrently collected with the MT inside the bay, suggesting that this is an improper method to estimate larval supply for this species. However, one caveat to these conclusions is that only daytime plankton net samples were assessed in this study. Additional studies also assessing nighttime plankton tow samples in relation to the other methods are warranted. Further, for species that show no association to
structurally complex habitats and that appear to avoid entering MTs, such as plaice larvae Pleuronectes platessa (H.W. unpubl. data), plankton net tows may be the best collection method available.
Artificial settlement substrates — temporary ‘settlement’ Shore crabs High losses of shore crab settlers were found on ASS that were immersed longer than 12 h, and the losses increased with increasing immersion time. Since the similar abundance of settlers collected on ASS provided with and without protective cages in Expt 3 indicate that predation was low on the ASS, these losses were most likely caused by emigration of postlarvae. Emigration by juvenile crabs is a less likely explanation for the loss, since benthic first-instar stages of shore crabs cannot swim and the ASS were suspended 3 m above the bottom. The metamorphosed megalopae might therefore become ‘stranded’ on the ASS, as indicated by the increasing proportion of first-instar crabs collected during the longer immersion times. This result suggests that most postlarvae ‘settle’ only temporarily on the ASS, and that the emigration rate is not constant but involves a delay of less than 24 h, after which a large proportion of the megalopae leaves the substrate. The number of shore crabs on the 24 h ASS reflected the number of crabs on the 12 h night ASS better than on the 12 h day ASS, suggesting that most megalopae that ‘settled’ during the day emigrated during the night. This is supported by a recent laboratory study, in which shore crab megalopae showed diurnal swimming behavior and emigrated from habitats at night (P.-O.M. unpubl. data). The lack of a constant difference between day and night may be explained by intense settlement during the 5 h of light in the morning before the night immersion was sampled. In an attempt to estimate the daily emigration rate from the ASS in the study, we fitted our data to a simple model that estimates the sum of settlers after 96 h for the different immersion times used: N 96 = [I × (96/IT )] + [I × (1 – E)] × [8 – (96/IT )] where N 96 = the sum of settlers after 96 h, I = total immigration to the ASS 12 h–1, IT = immersion time (h) and E = the proportional emigration occurring at night. This model assumes that a constant proportion of the megalopae that settled during the day emigrated only during their first night on the ASS. Employing a least-squares analysis and using an immigration rate of 30 (mean number of settlers on 12 h ASS), our data suggest that 83% of the megalopae left the ASS during the night. Thus, for
Moksnes & Wennhage: Estimating larval supply and settlement
shore crabs, it is inappropriate to refer to the transitional megalopae on ASS as ‘settled’, since this term refers to the termination of a pelagic larval phase and assumption of a benthic life (Scheltema 1974), often associated with metamorphosis. Instead, for this species, ASS appear to give a relative estimate of planktonic larval abundance of megalopae in a late development stage. Because the emigration rate from the ASS was fairly constant between days during this short experiment, an increase in larval supply over the 4 d period, found on the 12 h ASS, was detected also with the 24 and 48 h ASS. Thus, despite the high emigration rate, both 24 and 48 h ASS appeared to yield useful relative estimates of megalopal larval supply in this study. This was supported in Expt 3 in which the number of shore crabs collected on 24 h ASS inside the bay correlated significantly with the larval supply estimated by the MT-in, and reflected settlement densities found in the caged blue mussel habitat. A similar positive correlation between larval numbers on ASS, in plankton nets and in caged natural habitats have been found for Dungeness crabs on the US west coast where the arrival of larvae was highly predictable (Eggleston & Armstrong 1995). For estimates of larval supply of shore crabs in microtidal areas, we recommend that the ASS are fished for a maximum of 24 h and collected in the afternoon/ evening, since megalopae appear to emigrate at dusk (P.-O.M. unpubl. data), to minimize potential confounding effects of the megalopal emigration. In summary, the ASS appear to produce good integrated relative estimates of shore crab larval supply, and possibly also of initial settlement densities. Since the migration traps showed that recruitment of juvenile shore crabs to the assessed nursery area occurred almost entirely through planktonic megalopae, the estimate of settlement should be useful for recruitment studies of this species. However, the high proportion of transitional megalopae, indicated in this study, warrants further studies on how these estimates correlate with settlement densities in natural habitats. We urge colleagues using this technique with other decapod species to assess the proportion of transitional postlarvae on the ASS, and to assess how emigration rates from the collectors are affected by physical and biological factors. Such information would allow for the optimization of immersion and collection time, while assuring that estimates of larval abundance and settlement are not confounded by larval emigration.
271
96 h ASS was very similar to the number found on the 12 h ASS fished during the last night of the longer immersion times. Consequently, the number of larvae that ‘settled’ the day or days before was not reflected in the longer immersion times. Hence, the association to the ASS seems to be completely temporary, and after a delay of < 24 h, all shrimp appear to leave the ASS. Thus, ASS used for periods >12 h do not yield a useful integrated estimate of grass shrimp larval supply. In contrast to juvenile shore crabs, grass shrimp postlarvae and juveniles are competent swimmers that can easily migrate from the artificial substrate to more optimal habitats close by, which might explain why the ASS method produced poorer results for grass shrimp than for shore crabs.
Benthic-pelagic migration trap The new method using passive plankton net traps showed promising results in obtaining integrated migration estimates for both pelagic postlarvae and newly settled juvenile crabs and shrimp in shallow nursery areas. In comparison to plankton net tows and ASS, this method demonstrated several potential advantages: (1) It fished continuously and was therefore little affected by temporal variation in larval supply or current velocities. (2) It prevented escape and thus was not affected by variation in emigration rates. (3) It fished both plankton and animals moving on the sediment and could therefore also measure recruitment by migrating juveniles. (4) It fished in a directional manner, a known percentage of the water column at the opening of a bay, and could therefore potentially estimate the total net flux (number immigrating – number emigrating per unit time) of larvae and juveniles to a nursery area for the assessment of juvenile recruitment. Potential disadvantages with this technique are behavioral avoidance of the traps, clogging of traps by drifting macroalgae, and risk of predation inside the traps. In the present study, no drift algae were present in the water and predation on settlers inside the traps appeared to be limited since the juvenile crabs, shrimps and fishes found inside the MTs were too small (< 4 mm carapace width and length, and 15 mm total length, respectively) to be efficient predators on crab and shrimp recruits (Moksnes et al. 1998, P.-O.M. unpubl. data). The pattern of net fluxes and the role of juvenile migration for the recruitment were notably different for the 3 species investigated.
Grass shrimp Shore crabs Similar to the shore crabs, an increased immersion time resulted in an increasing loss of shrimp from the ASS. The number of shrimp found on both 24, 48 and
The MT appeared useful in obtaining integrated estimates of both shore crab larval supply and juvenile
272
Mar Ecol Prog Ser 209: 257–273, 2001
migration to a shallow nursery area. The number of settlers collected on MT-in correlated significantly with the abundance of settlers on ASS located inside the bay, and reflected changes in shore crab settlement-densities in caged mussel habitats. However, the results concerning net fluxes were less clear. During the first assessment of the MT, a significant net in-flux was detected daily over an 11 d period, even though emigration rates (estimated by MT-out) were high (on average, 45% of the number collected in MT-in). The second assessment occurred during a major peak in larval supply of shore crabs, and no significant difference was found between numbers of shore crabs collected with MT-in and MT-out. This result indicates very high emigration rates of megalopae, and that netrecruitment of juveniles was too small to be detected. However, the high numbers of first-instar crabs found in the caged blue mussel habitats indicate that high numbers of megalopae had settled in the bay during the study period. These conflicting results suggest that the settlement observed might only constitute a small fraction of the megalopae that passed through the bay, hidden from detection in the MT-data because of the large variance in larval abundance during the high settlement peak. It is also possible that emigration was unusually high during Expt 3 due to strong winds that continuously blew out of the bay during this period. Occasional sampling of the arms of the MTs during Expt 3 showed that high numbers of shore crab megalopae were clinging to the 800 µm net (the number of megalopae found inside the arms, which was not included in the analysis, constituted approximately 24% of the number in the trap on average). Thus, for this species, the MT appeared not to fish with 100% efficiency over the 2 m opening between the arms, and we suggest that the MT be used without arms for better estimates of shore crab megalopae fluxes in future studies. The significantly higher proportion of first-instar crabs found in the MTs compared to the other methods appears not to have been caused by higher metamorphosis rates among megalopae collected with MTs, since this was not indicated in the analysis of metamorphosis rates. The results indicate that high numbers of first instars entered the MTs (an estimated 40 crabs on average, constituting approximately 10% of the total number). The MT-in showed a significantly higher proportion of first instars than that collected with the MTout. This indicates that a small percentage (
